基于泛函网络的数值近似方法

NUMERICAL APPROXIMATION BASED ON A FUNCTIONAL NETWORK

根据泛函网络的网络结构模型及算法 ,给出一种近似二阶微分系统的解 ,采用泛函网络方法进行数值近似 ,得出了较好的结果 .实验研究表明 。

The network structure and algorithm of a functional network have been proposed recently. It is a natural generalization of neural networks. Good results are obtained for numerical approximation by using the functional network. It is shown by tests that a functional network is better than a BP network in numerical approximation.